KEMBAR78
Building Europe with Open LLMs - Manu Setälä.pptx
Mindtrek:
Building Europe
with Open LLMs
Manu Setälä, Head of Research
Solita Oy
About me and Solita Oy
Manu Setälä, Head of Research at Solita, https://www.linkedin.com/in/manuset/
• First software-related income in early 80's, with formal ICT studies starting in 1987.
• Interest in research is broad: public sector digital ecosystems, secure AI, AI-assisted solutions,
etc.
Solita Oy, Founded in 1996, https://www.solita.fi/
• 10 countries, 2200+ highly skilled Data, AI, Design, and Software Development professionals.
• We combine strategic consulting, service design, and software development to deliver impact
that lasts.
• Strong focus on AI: Leading the development and transition to Generative AI for our clients
(e.g., Solita FunctionAI platform - https://www.solita.fi/solita-function-ai/).
Background & why
• Europe is aging, and fewer people are required to provide services to society.
• More work will be done by ICT  Europe could become an AI-enabled society: public
services, infrastructure, and decision-making increasingly run on digital systems.
• This change is no longer a choice, but a necessity.
• In future LLMs play a significant role in the area of AI-enabled society.
• Significant part of society's activities are controlled by software
• As little manual work as possible
• (at least) Part of decision-making will be automated, hopefully safely, transparently, and responsibly.
• Dependency on non-European LLM providers (US, China) risks sovereignty, transparency, and
competitiveness.
• Open LLMs offer a European way: accessible, transparent, multilingual, and aligned with our
democratic values.
EU – same regulation 
same solutions
• Develop once, use in every country
• Open source
• European hosting / cloud
What (approach / solution)
• Open LLMs as building blocks: open-source models (Mistral, LLaMA, Falcon, etc.) enable
innovation, customization, and trust.
• Benefits:
• Transparency: bias and safety can be audited.
• Adaptability: models tailored to local languages and domains.
• Ecosystem growth: SMEs, researchers, and public sector can innovate without billion-dollar budgets.
• Challenges:
• Compute and infrastructure costs – this is both challenge and opportunity.
• Regulatory clarity (AI Act; Data Act, GDPR, …, liability in open models).
• Need for pan-European collaboration to reach scale!
• Not efficient that every country makes own solutions, and not very exportable  “Dev once use every country”
How (implementation / action)
• We must
• create a concept and reference models for an AI-enabled society;
• develop solutions that enable generative artificial intelligence to produce the necessary software
without traditional programming;
• establish the legal, technical, and organizational prerequisites for this new operating model.
The steps:
1. Exploration & pilots: national-level use cases in public services (regulation-driven)
2. European collaboration: Horizon-scale initiatives, open-source consortia, shared compute
infrastructure
3. Sustainable ecosystem: combining research, startups, large companies, and public sector
in an open innovation loop.
Strategic value (benefit / impact)
• Strengthens Europe’s technological sovereignty.
• Unlocks multilingual and multicultural AI leadership.
• Supports public trust and democratic legitimacy through openness.
• Builds a foundation for export-ready digital solutions and green scalability.
Some “calculations”
• Only public cloud costs in Europe about $142G 70% to USA ( = $100G = 10^11). CAGR 19%
• https://my.idc.com/getdoc.jsp?containerId=prEUR153716725
• According to IBISWorld software development in Europe is about 500 G€ - so with European
solutions this would be more than 20 % growth in Europe.
• Remember also license costs and so on – this could be 50% + 100% growth
• PLUS “snowball effect” and export
We can do it together
• Research in Universities and VTT e.g.
• Running: ANSE = AI Native Software Engineering, JKL/Tommi Mikkonen & TUNI/Pekka Abrahamsson
• Proposal preparation phase: Sovereign service infrastructure for digital transformation, JKL/Samuli Pekkola
• Many co-innovations already running: ELFMo, Genius, MAISA, …
• Public sector: potential for grant (Innovative public procurement – in FI: Innovatiiviset julkiset hankinnat)
Companies
R&D + execution
Public Body
user of new solution
• Tech competense
• Solution provider
Research
• Legal framework
• New Software Engineering
• Need & Cross-border solution steering
• Promoting legislative changes
Business Finland and Horizon Europe
Support for the organisations to try this new idea
University
research
Summary
“Building Europe with Open LLMs is not just about technology –
it is about sovereignty, values, and the ability to shape our own
digital future together.”
+
Huge potential business growth
MEGA = Make EUROPE Great Again
Discussion &
Questions
Manu Setälä
Solita Oy
Building Europe with Open LLMs - Manu Setälä.pptx

Building Europe with Open LLMs - Manu Setälä.pptx

  • 1.
    Mindtrek: Building Europe with OpenLLMs Manu Setälä, Head of Research Solita Oy
  • 2.
    About me andSolita Oy Manu Setälä, Head of Research at Solita, https://www.linkedin.com/in/manuset/ • First software-related income in early 80's, with formal ICT studies starting in 1987. • Interest in research is broad: public sector digital ecosystems, secure AI, AI-assisted solutions, etc. Solita Oy, Founded in 1996, https://www.solita.fi/ • 10 countries, 2200+ highly skilled Data, AI, Design, and Software Development professionals. • We combine strategic consulting, service design, and software development to deliver impact that lasts. • Strong focus on AI: Leading the development and transition to Generative AI for our clients (e.g., Solita FunctionAI platform - https://www.solita.fi/solita-function-ai/).
  • 3.
    Background & why •Europe is aging, and fewer people are required to provide services to society. • More work will be done by ICT  Europe could become an AI-enabled society: public services, infrastructure, and decision-making increasingly run on digital systems. • This change is no longer a choice, but a necessity. • In future LLMs play a significant role in the area of AI-enabled society. • Significant part of society's activities are controlled by software • As little manual work as possible • (at least) Part of decision-making will be automated, hopefully safely, transparently, and responsibly. • Dependency on non-European LLM providers (US, China) risks sovereignty, transparency, and competitiveness. • Open LLMs offer a European way: accessible, transparent, multilingual, and aligned with our democratic values.
  • 4.
    EU – sameregulation  same solutions • Develop once, use in every country • Open source • European hosting / cloud
  • 5.
    What (approach /solution) • Open LLMs as building blocks: open-source models (Mistral, LLaMA, Falcon, etc.) enable innovation, customization, and trust. • Benefits: • Transparency: bias and safety can be audited. • Adaptability: models tailored to local languages and domains. • Ecosystem growth: SMEs, researchers, and public sector can innovate without billion-dollar budgets. • Challenges: • Compute and infrastructure costs – this is both challenge and opportunity. • Regulatory clarity (AI Act; Data Act, GDPR, …, liability in open models). • Need for pan-European collaboration to reach scale! • Not efficient that every country makes own solutions, and not very exportable  “Dev once use every country”
  • 6.
    How (implementation /action) • We must • create a concept and reference models for an AI-enabled society; • develop solutions that enable generative artificial intelligence to produce the necessary software without traditional programming; • establish the legal, technical, and organizational prerequisites for this new operating model. The steps: 1. Exploration & pilots: national-level use cases in public services (regulation-driven) 2. European collaboration: Horizon-scale initiatives, open-source consortia, shared compute infrastructure 3. Sustainable ecosystem: combining research, startups, large companies, and public sector in an open innovation loop.
  • 7.
    Strategic value (benefit/ impact) • Strengthens Europe’s technological sovereignty. • Unlocks multilingual and multicultural AI leadership. • Supports public trust and democratic legitimacy through openness. • Builds a foundation for export-ready digital solutions and green scalability. Some “calculations” • Only public cloud costs in Europe about $142G 70% to USA ( = $100G = 10^11). CAGR 19% • https://my.idc.com/getdoc.jsp?containerId=prEUR153716725 • According to IBISWorld software development in Europe is about 500 G€ - so with European solutions this would be more than 20 % growth in Europe. • Remember also license costs and so on – this could be 50% + 100% growth • PLUS “snowball effect” and export
  • 8.
    We can doit together • Research in Universities and VTT e.g. • Running: ANSE = AI Native Software Engineering, JKL/Tommi Mikkonen & TUNI/Pekka Abrahamsson • Proposal preparation phase: Sovereign service infrastructure for digital transformation, JKL/Samuli Pekkola • Many co-innovations already running: ELFMo, Genius, MAISA, … • Public sector: potential for grant (Innovative public procurement – in FI: Innovatiiviset julkiset hankinnat) Companies R&D + execution Public Body user of new solution • Tech competense • Solution provider Research • Legal framework • New Software Engineering • Need & Cross-border solution steering • Promoting legislative changes Business Finland and Horizon Europe Support for the organisations to try this new idea University research
  • 9.
    Summary “Building Europe withOpen LLMs is not just about technology – it is about sovereignty, values, and the ability to shape our own digital future together.” + Huge potential business growth MEGA = Make EUROPE Great Again
  • 10.

Editor's Notes

  • #7 compound annual growth rate (CAGR)